Search This Blog


Tuesday, November 11, 2014

Understanding and Choices

One of the questions that people often ask is: do you tell Bow what to do or is that just what he chooses to do? For instance, when he is eating oatmeal, did I tell him to use the spoon?

No, I didn't tell him to use the spoon. I provided the spoon, and I told him to eat "nice". What does that mean? Mostly it means, don't make a mess. Be considerate of the fact that I have to clean up after you. But it does not mean you have to use Emily Post etiquette. It's the way I was raised. I didn't have to be perfect in my decorum, but I understood that I should not cause others more work than was necessary.

Everything with Bow is give and take. He does not always do exactly what I ask of him, but he takes reasonable suggestions under consideration. 

In trying to help me think of ways that I might prove what Bow knows, a colleague recently asked me if Bow would respond well to being rewarded for answering a question. We tried that more than once. He intentionally gave the wrong answer, every single time. It was not a random thing. It was better than chance! He was being obstructive because he resented the test,

So what was the reward we offered? It was chocolate. And he loves chocolate and does not get it all the time. But he wants to be given it as a treat, not a reward. He will do all sorts of nice and surprisingly considerate things, but not for a reward.

Bow is well fed. He has more than enough to eat. Yes, he does tell us what he wants to eat and when he is full. But he is not talking to us for a reward, and he would stop talking altogether, if he thought we were playing some kind of manipulative trick on him in our language use.

Well, why don't we just stop feeding him? He would probably be more cooperative then. Some people have suggested that. 

I would never do that. Withholding food is not ethical. It's wrong. And it would destroy the relationship of trust that I have built up with Bow. So all our language use remains in context. And as long as it's in context, then it's not a double blind test and will not pass muster for scientific purposes.

Well, how about if he were holding the hand of someone who did not speak his language? Then what he wrote would have to be coming from him, right? Remember, he only holds the hands of people he trusts. But, yes, something like that did happen once. The intern's name was Katie Thurston, And Bow did not entirely trust her, because she had a British accent. And he played a trick on her, by using a pun in Hebrew, and then just a spelled out Hebrew word, before he agreed to tell her in English that he was full. 

In Hebrew, the number seven is spelled שבע. But that is also how you spell the word for sated, meaning full.  He was being very tricky when he pointed at the 7, and of course Katie did not know what he meant. The next thing Bow spelled after the  7 to let Katie know he was full was מלא, which literally means full. But she did not know that Hebrew word.  Then, when she did not understand that, he spelled "because I am full," in answer to her question about why he did not want to eat.

Above is the transcript of that exchange.  It is not replicable, meaning that we can't get Bow to do that over and over again on command. It was something that he chose to do at the time, and like all real language use, it was a once in a lifetime exchange. You can't tell him: "Say that again." He won't.

Does Bow understand what we say to him? Yes. Does he obey commands? No, but he sometimes follows suggestions, when he himself decides it is a good idea. In all those cases, there is nothing context-free about it.

Real language use never is context free. It's a matter of give and take, a little cooperation and a little bargaining and making do with less than perfect results. 

No comments:

Post a Comment